What is the difference between partitions and replicas in Kafka?

In Kafka, partitions and replicas are different concepts related to scaling and fault tolerance.

Partitions are logical segments of a topic that allow data and load to be distributed among multiple brokers. Each partition is an ordered, immutable log of messages. Partitions provide parallelism: different consumers can read from different partitions simultaneously.

Replicas are copies of partitions stored on different brokers to ensure fault tolerance. One replica is the leader, and the others are followers. The leader handles all write and read operations, while followers synchronize data with the leader.

In summary:

• Partitions divide a topic into parts for scaling and parallel processing.
• Replicas provide redundancy and high data availability.

Example: a topic with 3 partitions and a replication factor of 2 will have 3 partitions, each stored on 2 brokers (1 leader + 1 replica).